www.acsnano.org
Peptide Brush Polymers and Nanoparticles with Enzyme-Regulated Structure and Charge for Inducing or Evading Macrophage Cell Uptake Lisa Adamiak,† Mollie A. Touve,†,‡,∥ Clare L. M. LeGuyader,† and Nathan C. Gianneschi*,†,‡,§,∥,⊥,# †
Department of Chemistry & Biochemistry, ‡Department of NanoEngineering, and §Materials Science & Engineering, University of California, San Diego, La Jolla, California 92093, United States ∥ Department of Chemistry, ⊥Department of Materials Science and Engineering, and #Department of Biomedical Engineering, Northwestern University, Evanston, Illinois 60208, United States S Supporting Information *
ABSTRACT: Cellular uptake by macrophages and ensuing clearance by the mononuclear phagocyte system stands as a significant biological barrier for nanoparticle therapeutics. While there is a growing body of work investigating the design principles essential for imparting nanomaterials with longcirculating characteristics and macrophage evasion, there is still a widespread need for examining stimuli-responsive systems, particularly well-characterized soft materials, which differ in their physiochemical properties prior to and after an applied stimulus. In this work, we describe the synthesis and formulation of polymeric nanoparticles (NPs) and soluble homopolymers (Ps) encoded with multiple copies of a peptide substrate for proteases. We examined the macrophage cell uptake of these materials, which vary in their peptide charge and conjugation (via the N- or C-terminus). Following treatment with a model protease, thermolysin, the NPs and Ps undergo changes in their morphology and charge. After proteolysis, zwitterionic NPs showed significant cellular uptake, with the C-terminus NP displaying higher internalization than its N-terminus analogue. Enzyme-cleaved homopolymers generally avoided assembly and uptake, though at higher concentrations, enzyme-cleaved N-terminus homopolymers assembled into discrete cylindrical structures, whereas Cterminus homopolymers remained dispersed. Overall, these studies highlight that maintaining control over NP and polymer design parameters can lead to well-defined biological responses. KEYWORDS: macrophage uptake, enzyme-responsive, stimuli-responsive, polymeric nanoparticle, self-assembly
M
deferred toxicity, ultimately preventing clinical translation of many nanomaterial-based therapeutics and diagnostics.5,6 To reduce NP susceptibility for opsonization and MPS clearance, many groups have developed strategies which consist of modifying NP surface chemistry,4,7−11 shape,12−14 size,15,16 or even surface topology.17−19 Though conjugation to poly(ethylene glycol) (PEG) represents the most commonly explored surface modification to render NPs resistant to opsonins,20,21 it is arguably unsuitable for bioactive NPs for which preservation of protein binding is required.22 Indeed, protein conjugates prepared using other types of hydrophilic
ost injectable nanoparticles (NPs) are known to suffer from premature removal from systemic circulation by the mononuclear phagocytic system (MPS). Macrophages of the MPS, typically Kupffer cells of the liver, bone marrow, and lung, recognize and sequester NPs through the adsorption of proteins to the nanomaterial surface, a process known as opsonization.1 The combined mechanisms of opsonization and macrophage recognition form the basis of immediate clearance via MPS (on the order of seconds) for injected NPs that are larger than the 10 nm renal filtration limit.2 For non-biodegradable NPs that cannot be destroyed by phagocytes, accumulation in MPS organs such as the liver and spleen is routinely observed.3,4 Long-term retention in MPS organs has raised concerns of heightened immune response or © 2017 American Chemical Society
Received: May 26, 2017 Accepted: September 15, 2017 Published: October 3, 2017 9877
DOI: 10.1021/acsnano.7b03686 ACS Nano 2017, 11, 9877−9888
Article
Cite This: ACS Nano 2017, 11, 9877-9888
Article
ACS Nano
Figure 1. Synthesis of (A) hydrophilic polymers, P, and (B) NPs formulated by dialysis of amphiphilic block copolymers. A library of polymers was synthesized utilizing zwitterionic, cationic, and anionic peptide monomers. The overall composition of the polymer structure was also tailored by modifying the nature of the peptide conjugation to the polymer (conjugation to the N-terminus or ε-amino group of a C-terminal lysine residue of the peptide). The proteolytic degradation of peptide side chains yields polymers with carboxylate or amine groups for N- and C-terminus materials, respectively.
NP sequestration in MPS organs. In general, stimuli-responsive systems have not been examined for their influence on macrophage cell recognition despite growing interest in their potential as drug carriers and diagnostics. Furthermore, previous studies have focused on a limited scope of functional groups presented at the NP surface. In particular, there are few examples detailing the effects of biosynthetic polymers on macrophage uptake.42 Given the encouraging potential of MMP-triggered therapeutic NP delivery to tumors and other inflamed tissues,36,37 we were interested in examining the effects of physiochemical differences before and after proteolysis on macrophage uptake. This active-targeting mechanism, which is enzyme-driven selfassembly, was also investigated for water-soluble polymers. Importantly, our general knowledge of enzyme-triggered assembly of initially fully dispersed polymers comprises only a handful of examples.43−45 Motivated by our efforts to develop protease-targeted diagnostic and drug delivery vehicles with favorable pharmacokinetic and biocompatible properties, specifically mitigation of off-target accumulation, we sought to evaluate how specific properties of the materials facilitate macrophage cell recognition and uptake, including surface charge, peptide composition, and nanoparticle versus homopolymer formulations as MMP-responsive scaffolds (Figure 1).
moieties, such as poly(zwitterions), exhibit similar biostabilizing properties but are markedly better at preserving protein bioactivity.23,24 Further, zwitterionic moieties have been shown to be nontoxic and nonimmune stimulatory.25 These strategies have been mostly disseminated for surface-functionalized gold NPs (AuNPs), in large part due to the ease of functionalization using alkanethiol ligand exchange reactions and well-established approaches for size and shape modifications.8,20,25−31 However, conclusions derived from studies of AuNPs need to be validated for soft materials within the size range suitable for in vivo applications.4,12 One key factor contributing to this discrepancy is the difficulty in precisely controlling the bottom-up fabrication of organic-based NPs that tend to vary in their size and surface chemistries.7 Thus, strategies to synthesize well-defined soft materials are necessary to evaluate MPS evasion and activation. Stimuli-responsive polymers are emerging as a promising class of biosynthetic materials, which can be produced through available synthetic methods to meet design specifications such as polymer size and chemical composition.32 Among stimuli, enzymes are unique in both selectivity and specificity for their substrates, along with their intrinsic capacity to act as molecular amplifiers.33,34 Enzymes operate under a range of mild, ambient solution conditions in vitro and, of course, physiological conditions, which is a feature that is desirable for biological applications.35 Recent work has established that enzyme-responsive NPs that display peptides on the NP surface allow for proteasetriggered accumulation in tumor tissue in vivo.36,37 The peptide moiety encompasses an optimized recognition sequence for matrix metalloproteinase-2 and -9 (MMP-2,-9), which are type 4 collagenases essential to basement membrane degradation processes necessary for angiogenesis or the formation of new blood vessels.38 As a result, MMPs are extensively involved in perpetual extracellular matrix remodeling within diseased tissues, which identifies them as key molecular targets for noninvasive NP treatments of myocardial infarction, hind limb ischemia, and cancer of all types.39−41 Though active-targeting strategies such as protease-triggered self-assembly have improved NP accumulation and retention in tumors or areas of infarcts, there remains a prominent issue of rapid, off-target
RESULTS AND DISCUSSION Design of NPs and Soluble Homopolymers. For this work, we designed NPs and homopolymers (Ps) that are susceptible to cleavage by MMP-9, which is a known inflammation-associated disease biomarker.38 To generate the library of materials used in this study, norbornyl peptides were polymerized either as a discrete block in amphiphilic block copolymers or as homopolymers via ring-opening metathesis polymerization (ROMP) using a pyridyl-modified third generation Grubbs’ initiator, [Ru] (Figure 1). Given the tolerance of this initiator to functionally complex amino acids, peptide polymers can readily be prepared using a graf t-through approach, enabling high-density incorporation of peptide side chains, circumventing the need for postpolymerization modification steps, and providing a living polymerization for control over degree of polymerization (DP).46−48 In order to 9878
DOI: 10.1021/acsnano.7b03686 ACS Nano 2017, 11, 9877−9888
Article
ACS Nano enable visualization of the materials in cells, a fluorophore was used to end-label the polymers by way of a fluorescein chain transfer agent (CTA) added to the living polymer chain.49 Using standard synthetic procedures, we designed six peptides with zwitterionic, cationic, or anionic charge. The amino acid sequence (GPLGLAG) was conserved across NPs and Ps because it is an optimized recognition sequence for MMP-9 (Figures 1 and S1).50 Conjugation of a norbornene(Nhexanoic acid) moiety to either the N-terminus or to the εamino group of a C-terminal lysine residue afforded two structural isomers, which upon enzyme cleavage rendered two possible norbornyl peptide products containing either (1) carboxylic acids for N-terminus conjugates or (2) primary amines for C-terminus conjugates (Figure 1). This in turn allowed for the evaluation of cell uptake of NPs and Ps as a function of morphology and the charge borne by the material before and after enzyme cleavage. Uniform, spherical NPs approximately 20 nm in diameter were formulated by slow dialysis of amphiphilic block copolymers into Dulbecco’s phosphate buffered saline (DPBS), as measured by dry-state transmission electron microscopy (TEM) (Figure 2) and dynamic light scattering
Obtained values were in good agreement with theoretical values expected from the peptide charge contribution with the exception of zwitterionic NP1, which exhibited a negative value (Table 1). Insufficient signals were measured by DLS for water-soluble homopolymers in DPBS (Figure S7), confirming that they are dispersed polymer solutions. Cellular Uptake of NPs and Homopolymers. We selected the murine macrophage cell line RAW 264.7 as a model cell system because it is well-documented that macrophages remove most of the administered dose of NPs in vivo.3 Fluorescence-based assays using RAW 264.7 cells were used to evaluate the cell uptake of NPs, hydrophilic Ps, and their enzymatically degraded products. Our first objective was to examine how prepackaging peptides of different charges and compositions as NP or soluble polymer scaffolds provokes cell entry. To this end, the cellular uptake of materials by RAW 264.7 cells was assessed by flow cytometry and was quantified as the normalized mean fluorescence, which was calculated as the ratio of mean fluorescence count of cells treated with material to the mean fluorescence count of cells treated with vehicle (DPBS). From these experiments, cells treated with cationic NP2 and NP5 displayed the greatest fluorescence signal intensities compared to all other materials (Figure 3). Positively
Figure 3. Cell uptake of (A) NPs and (B) soluble homopolymers (Ps) at 1 μM after 3 h incubation with RAW 264.7 cells.
Figure 2. Dry-state TEM images of spherical NPs of various charges formulated from amphiphilic block copolymers by dialysis from organic cosolvents into DPBS.
charged NPs are generally known to elicit greater cell internalization relative to other NPs with negative surface charge.52 Interestingly, cells treated with anionic NP3 showed a 4-fold enhancement relative to cells treated with its C-terminal counterpart (NP6), which showed negligible internalization (Figure 3A). Although these systems are similar in size and ζpotential values, this discrepancy in cellular response may arise from differences in the arrangement of anionic peptides on the NP surface invoked by the N- or C-terminus connectivity. In contrast to the other charged NPs, zwitterionic NP1 and NP4 provoked no cellular response (Figure 3A). Interestingly, zwitterionic NPs formulated from random copolymers (NP7) were also resistant to cellular uptake, suggesting that complex polymer architectures may evade macrophage uptake by virtue of charge alone (Figures S9 and S10; see Supporting Information). These findings are in agreement with the stealth-like properties imparted by zwitterionic NP surface coatings and may offer an attractive approach to the synthesis of NPs formulated from block copolymers.53,54 All peptides incorporated as soluble homopolymers exhibited negligible cell internalization (Figure 3B). Overall, the data indicate that NPs bearing cationic or anionic peptides led to their cell uptake, whereas NPs bearing zwitterionic peptides
(DLS) (Table 1 and Figure S6). To minimize variations in NP size, the DP of both discrete blocks was conserved across the library. Further, a relatively high wt % of the hydrophilic peptide block was maintained to restrain the surface curvature and ensure the assembly of spherical micelles.51 ζ-Potentials were measured for NPs after dialysis from DPBS into a low ionic strength buffer (10 mM phosphate buffer, pH 7.5). Table 1. DLS and ζ-Potential Measurements of NPs NP
polymera
Dh (nm)b
NP1 NP2 NP3 NP4 NP5 NP6
Ph41-b-126 Ph41-b-232 Ph41-b-325 Ph41-b-415 Ph41-b-527 Ph41-b-623
22.4 15.6 18.9 9.9 15.3 17.1
ζ (mV)c −20.5 +13.4 −19.7 +2.90 +16.7 −17.8
± ± ± ± ± ±
2.0 0.9 1.3 0.5 1.2 1.3
a
Block copolymer used to formulate NPs. bHydrodynamic diameter measured by DLS. cNPs measured in 10 mM phosphate buffer (pH 7.5). Average of three measurements. 9879
DOI: 10.1021/acsnano.7b03686 ACS Nano 2017, 11, 9877−9888
Article
ACS Nano
serum and serum-free conditions.63 The serum-dependent uptake of anionic NPs by phagocytic cells has been observed in previous studies,64,65 though in these cases (50 and 100 nm iron oxide NPs or carboxylate-modified microspheres of 20 and 200 nm), the presence of serum increased the cytotoxicity and cell uptake of the material, relative to serum-free conditions. For the cell uptake of carboxylate-coated NPs by nonphagocytic cells (HeLa and MDCK), it was shown that preincubation of NPs with serum reduces NP uptake.66 In these examples, it was speculated that serum directly impacts the ability of NPs to interact with the cell surface, possibly due to reversible adsorption and a decrease in the NP ζ-potential, despite no change in NP size by DLS. It is possible that a similar mechanism is at play with cationic NP2 and NP5. In general, nonphagocytic cells preferentially internalize cationic NPs, whereas phagocytic cells take up anionic NPs, although there is a size dependence.67 Intriguingly, the flow cytometry data indicate the reverse trend for the library presented here. To determine whether enhanced cell uptake for cationic NPs in serum-free conditions was unique to the NP scaffold, the serum-free cell uptake of homopolymers was also examined (Figure S11). There was no difference in the cell uptake between serum and serum-free conditions, indicating potentially different mechanisms of serum adsorption and cell− material interactions for soluble homopolymers. Enzyme-Triggered Assembly and Cell Uptake of NPs and Homopolymers. Our next objective was to determine how enzyme cleavage of peptides displayed by NPs and Ps incites differences in their cell uptake. Enzymatic stimuli have been used by others as an activatable cell internalization mechanism, though the effects on macrophage recognition have not been examined.68−70 It could be envisioned that an enzyme-targeting strategy for macrophages may be useful for promoting macrophage differentiation, for applications in activating M2 phenotypes, also referred to as tumor-associated macrophages (TAMs), to “eat” cancer.71−76 In this regard, using enzymes to trigger a morphology switch in NPs and Ps to form micrometer-sized clusters could lead to enhanced uptake by phagocytic cells, which have been shown to preferentially internalize particles between 2 and 3 μm.77 We also hypothesized that the exposure of carboxylates for N-terminus conjugates and amines for C-terminus conjugates by enzymatic hydrolysis could affect the cellular uptake of these materials in a different manner, given the previous assessment of chargedependent cell uptake of NPs (Figure 3A). For example, Cterminus conjugates may be internalized to a greater extent than N-terminus conjugates after proteolysis as the materials are enzymatically converted to cationic products. For these experiments, NPs and Ps were treated with a model enzyme, thermolysin, which is a highly active bacterial zinc protease.78 Thermolysin was used as an in vitro proxy for MMP-9 because it is easily available, has high activity, and can potentially cleave at two sites adjacent to the leucine residues in position P1′, which results in discriminate consumption of the peptide. NPs and Ps were subjected to thermolysin for 15 h at 37 °C in 50 mM Tris pH 7.4, after which the enzyme was chemically denatured with 10% v/v 0.5 mM ethylenediaminetetraacetic acid (EDTA). Controls were prepared in the same fashion except that thermolysin was chemically denatured prior to incubation with the materials. The enzyme-cleaved materials were then incubated with RAW 264.7 cells to evaluate their cellular uptake. The solutions were also analyzed by reversephase high-performance liquid chromatography (RP-HPLC)
avoided cell uptake. Further, soluble polymers, regardless of charge, may generally evade macrophage uptake. Energy- and Serum-Dependent Cell Uptake of NPs. To probe whether the mechanism of cellular uptake of NPs involves passive diffusion across the cellular membrane or receptor-mediated internalization (an active, energy-dependent process), flow cytometry experiments were repeated at low temperature (4 °C) (Figure 4). At low temperatures, surface-
Figure 4. Mechanistic studies of cell uptake of NPs (zwitterionic NP1, NP4; cationic NP2, NP5; and anionic NP3, NP6) at 1 μM by RAW 264.7 cells after 3 h incubation. Cells were treated at 37 °C with NPs in the presence of serum (heat-denatured FBS or competent FBS) or without serum (no FBS) to assess serum dependence on cell uptake. An active mechanism for cell uptake was assessed by treating the cells with NPs at 4 °C to arrest cell surface receptor activity; * denotes p < 0.05, and ** denotes p < 0.01.
receptor-mediated cell uptake has been shown to be dramatically reduced.55,56 All NPs (predominantly cationic and anionic NPs) showed suppressed cell uptake at 4 °C, which suggests that the internalization mechanism may be through a receptor-mediated process (Figure 4). Having previously established the roles of charge on the cellular uptake of NPs in complete cell culture medium containing 10% heatinactivated fetal bovine serum (FBS), the cellular uptake of these materials was then assessed using serum-free and competent serum conditions to investigate the effects of serum on cell internalization. Studies on the cellular uptake of a variety of spherical NPs in biological environments show that the protein corona plays a central role in the resulting cellular− NP interactions.8,26,29,57−59 Competent FBS was used because it contains complement proteins (not heat inactivated) that affect opsonin−macrophage recognition processes.16,60−62 In these experiments, RAW 264.7 cells were treated with the materials for 3 h in the presence of competent serum or no serum. The flow cytometry results of cationic NPs (NP2 and NP5) showed a significant increase in cell uptake in the absence of serum (Figure 4). There was no difference in the uptake of all NPs between competent serum and heat-inactivated serum conditions. Notably, zwitterionic NP1 and NP4 showed no change in cell uptake, which may suggest the absence of a hard corona for these charged NPs. This is consistent with other studies; for example, Rotello and co-workers observed that the internalization of zwitterionic AuNPs was the same in both 9880
DOI: 10.1021/acsnano.7b03686 ACS Nano 2017, 11, 9877−9888
Article
ACS Nano
Figure 5. Cell uptake of NPs after thermolysin treatment for 15 h at 37 °C by RAW 264.7 cells. (A) Cell uptake was quantified by flow cytometry of cells incubated with NPs following treatment with active and denatured enzyme. (B) Representative histograms of the gated cell populations giving rise to fluorescence intensity for NPs. (C) Percent cleavage of NPs used for cell assays quantified by RP-HPLC (see Figure S18 for chromatograms). (D) Representative dry-state, uranyl-acetate-stained TEM images of NP1 treated with denatured enzyme (left panel) or active enzyme (right panel). Yellow arrows indicate intact particles; * denotes p < 0.05, and ** denotes p < 0.01; ns denotes not significant, and nd denotes not determined: note that only a small percentage (∼5%) of the gated cell population contributes to the large values of mean fluorescence intensity of cells treated with NP6.
fold increase in mean fluorescence counts compared to that of its N-terminus counterpart, NP1 (Figure 5A). This finding is consistent with the previous experiment showing that cationic NPs elicited greater cellular uptake by RAW 264.7 cells, though, in this case, the comparison is between micrometer-sized entities. Nevertheless, the cellular uptake of cationic NP2 and NP5 was not significantly different between denatured and active thermolysin conditions (Figure 5A). Interestingly, NP6 displayed the largest mean fluorescence intensity after enzyme treatment. However, only a small percentage (∼5% for NP6) of the gated cell population contributed to the large fluorescence intensity (Figure 5B). Exclusive to zwitterionic NP1 and NP4, enzyme cleavage was able to activate their cell internalization, suggesting that the nature of the charge switch may be important. In contrast to NPs, enzyme processing of homopolymers generally elicited no change in their cell uptake (Figure 6). After enzyme treatment, Ps still appeared dispersed in solution, though some smaller nanostructures were observed by TEM, particularly for N-terminus conjugates, P1 and P3 (Figures 6D and S21). Interestingly, only zwitterionic P1 showed a significant difference in mean fluorescence intensity after enzyme treatment compared to the denatured enzyme control; however, this enhancement was a marginal increase (Figure 6A). Similar to NP6, anionic P6 also showed a large increase in cellular uptake after enzyme treatment, though only ∼10% of the gated cell population contributed to the fluorescence intensity (Figure 6B).
and TEM to quantify the amount of converted peptide substrate and to examine changes in material morphology, respectively. Enzyme cleavage of NPs by thermolysin in all cases resulted in the formation of heterogeneous micrometersized aggregates, consistent with previous reports of enzymetriggered assembly of NPs (Figure 5 and Figures S18 and S20).37,79 Control solutions of NPs treated with denatured thermolysin show that the spherical NP morphologies remain intact in the absence of enzyme cleavage (Figures 5D and S20). Nevertheless, a significant portion of NP1 and NP3 remained as discrete, spherical NPs, presumably due to the incomplete conversion of all peptide substrates (Figures 5C and S20). Interestingly, C-terminus NPs (NP4, NP5, NP6) outperformed N-terminus NPs (NP1, NP2, NP3) with respect to enzyme conversion, with the exception of cationic analogues, which were both cleaved quantitatively (Figure 5C). This is likely due to the extension of the peptide substrate farther from the polymer backbone in the C-terminus conjugates (Figure S1). This observation is also in agreement with a previous report by Wooley and co-workers, wherein positively charged NPs were shown to attract enzymes bearing a net negative charge under physiological conditions.80 Thermolysin has an isoelectric point of 4.45 and should, in principle, exhibit strong electrostatic interactions with cationic substrates, leading to higher rates of hydrolysis for these materials.78 C-terminus NPs also showed higher cell uptake than Nterminus NPs after enzyme processing in some cases. For example, enzyme-cleaved zwitterionic NP4 showed at least a 49881
DOI: 10.1021/acsnano.7b03686 ACS Nano 2017, 11, 9877−9888
Article
ACS Nano
Figure 6. Cell uptake of Ps after thermolysin treatment for 15 h at 37 °C by RAW 264.7 cells. (A) Cell uptake was quantified by flow cytometry of cells treated with Ps following treatment with active and denatured enzyme. (B) Representative histograms of the gated cell populations giving rise to fluorescence intensity for Ps. (C) Percent cleavage of Ps used for cell assays as quantified by RP-HPLC (see Figure S19 for chromatograms). (D) Representative dry-state, uranyl-acetate-stained TEM images of P1 treated with denatured enzyme (left panel) or active enzyme (right panel). Red arrow indicates the presence of small aggregates after enzyme cleavage; ** denotes p < 0.01; ns denotes not significant, and nd denotes not determined: note that only a small percentage (∼10%) of the gated cell population contributes to the large values of mean fluorescence intensity of cells treated with P6.
(Figures 5A and 6A), albeit a small percentage of the gated population, may be consistent with their high cytotoxicity. Specifically, C-terminus conjugates, such as NP6 and P6, are enzymatically cleaved to form positively charged amines under physiological pH (Figure 1). The cationic charge of their enzyme-cleaved products may induce their high cell uptake and consequently high cytotoxicity.67 It is unclear why other C-terminus conjugates (NP4, NP5 and P4, P5) do not follow the same pattern. This observation indicates that the nature of the charge switch of the material due to proteolysis may affect the cell uptake of the material and potentially its cytotoxicity. Enzyme-Triggered Assembly of Homopolymers at Higher Concentrations. Noting the discrepancy between enzyme-triggered cell uptake of P1 but not other Ps, we assessed whether higher concentrations of Ps would afford robust assemblies, similar to the ones observed for NPs. Increasing enzyme concentration has been demonstrated as an accessible strategy to induce kinetic control over supramolecular assembly of dipeptide precursors.81 To this end, we synthesized zwitterionic polymers, N-terminus P1 (DP 18) and C-terminus P4 (DP25), to examine how proteolysis of Ps varies with peptide composition. Note these polymerizations were terminated with ethyl vinyl ether and do not contain a fluorescein end group. P1 (DP 18) and P4 (DP 25) at 1 mM were incubated with 1 μM thermolysin at 37 °C for 24 h in 50 mM Tris pH 7.4 buffer. Controls were prepared in a similar fashion as described earlier. Remarkably, discrete cylinder-like
To identify how enzyme cleavage of the materials may impact their cytotoxicity, the CellTiter-Blue assay was used to assess the cell viability of RAW 264.7 cells after 24 h incubation with NPs and Ps treated with and without thermolysin (Figure 7). Before enzyme treatment, nearly all treatments exhibited >80% cell viability, with the exception of NP5, which exhibited 65% viability over the 24 h period (Figure 7A). The enzymecleaved NPs and Ps were also not sufficiently cytotoxic except for anionic, C-terminus NP6 and P6 (Figure 7). The large cell uptake of C-terminus, anionic materials after enzyme cleavage
Figure 7. Cell viability of RAW 264.7 cells after incubation with NPs and Ps before and after enzyme cleavage. 9882
DOI: 10.1021/acsnano.7b03686 ACS Nano 2017, 11, 9877−9888
Article
ACS Nano
Figure 8. Enzyme-triggered assembly of N-terminus homopolymers (1 mM) by 1 μM thermolysin at 37 °C. (A) N-terminus P1 (DP 18) and Cterminus P4 (DP 25) were treated with thermolysin for 24 h. The cleavage sites are depicted by the red and blue arrows, respectively. Drystate TEM (middle panel) indicates the presence of cylindrical micelles after enzyme cleavage of P1 (DP 18) but minimal residue after enzyme cleavage of P4 (DP 25). RP-HPLC chromatograms (bottom panel) show the expected cleavage products. (B) Percent conversion of P1 (DP 18). The black arrow indicates the onset of cylindrical micelle formation as shown by dry-state TEM in (C). Cleavage of monomer 1 is shown for comparison. (D) Assembly of the authentic cleaved version of P1 (truncated P1) after dialysis into Tris pH 7.4 buffer from DMSO (see Supporting Information).
cleaved P1, was then formulated into nanostructures by slow dialysis from organic cosolvents into aqueous buffer (50 mM Tris, pH 7.4). After slow transition into buffer, truncated P1 formed predominantly aggregate structures by dry-state TEM (Figures 8D and S27). Although cylindrical structures were identified after dialysis from dimethylsulfoxide (DMSO), they were fairly nondiscrete and unlike the ones observed in enzymatic reactions, which provides additional evidence that these structures are kinetically formed. At present, it is difficult to deconvolute the specific contributions leading to this phenomenon, as it is well-known that small perturbations in polymer composition and structure can largely influence unimer exchange rates and dictate assembly in unpredictable ways.51 Moreover, the dynamic nature of the relationship between enzyme cleavage and assembly (occurring on the time scale of minutes to hours) complicates matters further by changing the composition and the hydrophobic balance simultaneously. Further investigations in our laboratory are needed to verify the mechanisms that contribute to assembly of N-terminus polymers. Nonetheless, the principles learned here could find utility in the development of soluble polymers as previously untapped scaffolds for enzyme-directed assembly and as a strategy for evading or activating macrophage uptake.
structures were observed for N-terminus P1 (DP 18) postenzyme treatment by dry-state TEM and cryo-EM, whereas there was no evidence for assembly formation for C-terminus P4 (DP 25) (Figures 8 and S23). RP-HPLC was used to quantify the amount of converted product over time as done previously (Figure 8A). Although a small amount of substrate was converted (∼3%) in the control solution for P1 (DP 18), there was no observable assembly formation by TEM, indicating that the active enzyme is necessary for self-assembly to occur (Figures 8A and S23A). A survey of other Ps at various DPs, including cationic P2 and P5, indicated that enzymedirected self-assembly is a general feature for N-terminus Ps but not for C-terminus ones (Figures S23B and S24). We further sought to delineate whether the enzyme-triggered assembly of N-terminus polymers into cylindrical structures was the product of a thermodynamically or kinetically driven solution-phase process. Analysis of a solution of P1 (DP 18) with thermolysin at various time points (20 min, 40 min, 1 h, and 24 h) by TEM and DLS indicated that cylindrical micelles emerged early in the reaction (Figures 8C and S25) and at approximately 50% conversion of substrates (Figure 8B), suggesting that these assemblies are kinetically derived. By 24 h, a mixture of phases, specifically aggregates on the scale of several hundred nanometers along with cylindrical structures, was observed (Figure S25). DLS indicated the presence of bimodal populations from the onset of the reaction, though by 24 h, larger structures (>100 nm) were predominant (Figure S25). In an effort to reproduce the cylindrical phases of enzymatically cleaved P1, an authentically synthesized “truncated” P1 mimic was prepared (Figures 8D and S26). Truncated P1 (DP 22), which is nearly identical to the fully
CONCLUSIONS In summary, we have shown that by utilizing ROMP we can tailor NPs and peptide brush polymers (Ps) to evade macrophages or activate their cellular uptake by enzymetriggered assembly. Given the exceptionally narrow range in NP diameter (∼20 nm), these materials were good candidates to assess the effect of NP and P charge, as well as peptide 9883
DOI: 10.1021/acsnano.7b03686 ACS Nano 2017, 11, 9877−9888
Article
ACS Nano
(Biorad, cat. #7321010) for 5 min, followed by draining, rinsing one time with DMF, and then applying fresh methylpiperidine solution for another 10 min. After thorough rinsing with DMF following deprotection steps, amide coupling reactions proceeded for a minimum of 45 min using 3 equiv of FMOC-protected amino acids (AAs), 2.9 equiv of HBTU, and 6 equiv of DIPEA. FMOC-Lys(Mtt)OH residues (AAPPTec, cat. #AFK125) (1.5 equiv) were doublecoupled (i.e., subjected to two consecutive applications of fresh AA/ HBTU/DIPEA solutions) at the first conjugation step. Peptide monomers were prepared by double-coupling to 1.5 equiv N(hexanoic acid)-cis-5-norbornene-exo-dicarboximide at the N-terminus of the peptide or the ε-amino group of a lysine residue near the Cterminus. (3-Carboxypropyl)trimethylammonium chloride was purchased from Sigma (403245) and double-coupled (1.5 equiv) at the N- or C-terminus of the peptide in a similar manner to generate zwitterionic and cationic peptides. Acetylation of the N- or C-terminus lysine residue afforded anionic peptides. Selective deprotection of methyltrityl (Mtt) protecting groups on resin was afforded by shaking the resin five times in TFA/triisopropylsilane (TIPS)/DCM (3:5:92 v/v/v) (approximately 6 mL per gram resin) for 7 min each, followed by rinsing with DCM. Full deprotection was confirmed via the Kaiser test. Peptides were cleaved from the resin for 1 h using a solution of TFA/TIPS/water (95:2.5:2.5 v/v/v). The TFA solution was drained into a conical tube and evaporated. The concentrated peptide solution was then precipitated in cold ether and centrifuged, and the pellet was purified by RP-HPLC. The identity and purity of each peptide were verified via ESI-MS analysis and the presence of a single peak in the analytical RP-HPLC chromatogram. Polymerizations. All polymerizations were performed in a glovebox under N2 gas using dry, deoxygenated DMF obtained by the freeze−pump−thaw method. Polymerizations that have not been reported previously in the literature were performed in DMF-d7 in J Young NMR tubes and assessed via 1H NMR to confirm the complete consumption of monomer and determine the reaction time required for completion. A typical procedure to synthesize a homopolymer of DP 25 involves adding 50 μL of [Ru] in anhydrous DMF (0.5 mmol, 1 equiv) to a solution of peptide monomer (10 mmol, 20 equiv) in DMF to a final volume of 450 μL. The solution was shaken to ensure mixing. End functionalization was achieved by adding a fluorescein CTA (1.5 equiv) for 3 h as described previously followed by the addition of excess ethyl vinyl ether to ensure complete termination.49 An aliquot (15 μL) was removed to characterize the polymer molecular weight and dispersity (Đ or Mw/Mn) via SEC-MALS. The homopolymer was precipitated using cold diethyl ether and centrifuged at 3000 rpm for 7 min. The ether was decanted, and the pellet was triturated two times with DMF and cold ether, followed by centrifugation. The remaining pellet was dissolved in water with a minimal amount of acetonitrile and lyophilized to afford a white powder. Block copolymers were prepared by polymerizing norbornyl phenyl (40 equiv) to completion, splitting the reaction into six vials, then adding the appropriate peptide monomer as the second block (20 equiv), and finally end-labeling with fluorescein and terminating the polymerization as described above. Purification was carried out as described in the homopolymer procedure. Fluorescein-containing polymers were treated with aqueuos NH4OH for 30 min to remove the pivalate protecting groups. All polymers were characterized by SEC-MALS in DMF. Nanoparticle Formulation. Assembly of amphiphilic block copolymers into spherical NPs was achieved by first dissolving the amphiphile in an organic solvent (DMF, acetonitrile, or DMSO). Milli-Q water was added dropwise and the solution mixed by hand until 10% v/v H2O was reached (1 mg/mL polymer). The solutions were equilibrated for approximately 3 h prior to dialysis. The solutions were transferred to a 3500 MWCO snakeskin dialysis tube (ThermoFischer, #68035) and dialyzed against 1 L of Milli-Q water. The water was refreshed after 8 h. This process was repeated once more, and then a few aliquots were removed and lyophilized in tared vials. The remaining suspension was concentrated with a 10 000 MWCO centrifugal filter (Millipore) and analyzed for the polymer concentration using a standard curve generated from the dried aliquots.
composition, on their cellular uptake by RAW 264.7 cells using flow cytometry. All hydrophilic Ps (DP 25) were able to sufficiently thwart macrophage recognition and uptake. This is in contrast to NP analogues, wherein cationic and anionic versions displayed higher levels of cellular uptake compared to zwitterionic ones. We then examined whether enzyme processing of peptide side chains of NPs and Ps could lead to enhanced macrophage cell uptake as a result of changes in morphology and charge of the materials. Zwitterionic NPs were able to undergo cellular internalization by RAW 264.7 cells after proteolytic cleavage, whereas other NP analogues generally showed minimal changes, despite also undergoing assembly to form micrometer-sized aggregates. These results indicate that the charge switch of the material after proteolysis was not necessarily the prevailing factor for cell uptake but must also be coupled with the morphological change. In contrast, only zwitterionic P1 and anionic P3 formed assemblies, albeit much smaller in size, and showed minimal uptake after proteolysis. At higher concentrations, however, enzyme cleavage of N-terminus Ps resulted in their assembly into fairly discrete cylindrical micelles, whereas C-terminus Ps remained dispersed. This is particularly significant, given that there are few reports outlining the enzyme-directed assembly of initially non-assembled polymers.43−45 Overall, the results indicate that incorporation of functional peptides into soluble polymers may be a general strategy for reducing MPS accumulation in vivo. Further, an enzyme-responsive strategy may be employed to activate macrophage cell uptake of zwitterionic NPs and potentially Ps.
MATERIALS AND METHODS General Materials. Amino acids used in solid-phase peptide synthesis (SPPS) were purchased from AAPPTec and Novabiochem. All other materials were purchased from Sigma-Aldrich and used without purification unless otherwise noted. ((H 2 IMES)(pyr)2(Cl)2RuCHPh) was prepared as previously described.82 Synthesis of the phenyl norbornyl monomer (N-benzyl)-5-norbornene-exo-2,3-dicarboximide was previously published.49 Analytical RPHPLC was performed using a Jupiter Proteo 90A Phenomenex column (150 × 4.6 mm) using a Hitachi-Elite LaChrom L-2130 pump with a UV−Vis detector (Hitachi-Elite LaChrom L-2420) monitoring at 214 nm. Peptides were purified on a Jupiter Proteo 90A Phenomenex column (2050 × 25.0 mm) on a Waters DeltaPrep 300 system. A gradient of 0.1% TFA in water (buffer A) and 0.1% TFA in acetonitrile (buffer B) was used for RP-HPLC analyses and peptide purifications. Size-exclusion chromatography was performed using Phenomenex Phenogel 5 u 10, 1−75 K, 300 × 7.80 mm in series with Phenomenex Phenogel 5 u 10, 10−1000 K, 300 × 7.80 mm using a Shimadzu pump equipped with a multiangle light scattering detector (DAWN-HELIO, Wyatt Technology) and a refractive index detector (HITACHI L2490 or a Wyatt Optilab T-rEX detector) normalized to a 30 000 MW polystyrene standard. The eluent used was 0.05 M LiBr in dimethylformamide (DMF) at a flow rate of 0.75 mL/min. DLS measurements were obtained using a DynaPro NanoStar (Wyatt Technologies). 1H NMR spectra were recorded on a 500 MHz Varian VX spectrometer. Chemical shifts are reported in parts per million relative to residual solvent peaks. Peptide Synthesis. Peptides were generated by the incorporation (zwitterionic 1, 3, and anionic 4, 6) or omission (cationic 2, 5) of an anionic glutamate residue in the amino acid sequence. Peptides were synthesized using standard FMOC SPPS procedures on an AAPPTec Focus XC automated synthesizer. Peptides were prepared on Rink amide MBHA (AAPPTec, cat. #RRZ005), Wang-Gly (AAPPTec, cat. #RWG101), or Wang (AAPPTec, cat. #RWZ001) resins. Briefly, FMOC deprotection was achieved by shaking peptide-bound resin with 20% methylpiperidine/DMF in a plastic chromatography vessel 9884
DOI: 10.1021/acsnano.7b03686 ACS Nano 2017, 11, 9877−9888
Article
ACS Nano
Affymetrix, cat. #16920), and finally rinsed with DPBS (0.5 mL). The cells were lifted from the plate using 0.25 mL of 0.25% trypsin with EDTA (GIBCO Life Tech., cat. #15090-046) for 15 min. The trypsin solution was pipetted up and down several times to dislodge the cells and transferred to Eppendorf tubes. Fresh DMEM (0.45 mL) was added to the wells to collect any remaining cells and transferred to the Eppendorf tubes. Finally, DPBS (0.7 mL) was added to the tubes, and the suspensions were centrifuged. After the supernatant was aspirated, cell pellets were suspended in 60 μL of DPBS and stored on ice prior to flow cytometry measurements. Fluorescence activated cell sorting (FACS) data (10 000 events on three separate cell cultures) were acquired on an Accuri C6 flow cytometer set to default “3 blue 1 red” configuration with standard optics and slow fluidics (14 μL/min). Data are reported as the normalized mean fluorescence, which is the ratio of mean fluorescence intensity of cells treated with material to cells treated with vehicle (DPBS). Data were analyzed using FlowJo software. Experiments were performed three times on three separate subcultures. Standard error is plotted for all experiments unless otherwise indicated. Mechanistic Studies by Flow Cytometry. For mechanistic studies, cells were plated and treated as described above. For studies at reduced temperature, cells were incubated at 4 °C immediately following the addition of the treatment and during incubation. For studies without FBS, cells were treated with material dissolved in DMEM supplemented with all the components as described except for heat-inactivated FBS. For studies in competent FBS, cells were treated with material dissolved in DMEM supplemented with all the components as described except that heat-inactivated FBS was replaced with non-heat-inactivated FBS (Omega, FB-01). Cellular Uptake of Enzyme-Cleaved Materials by Flow Cytometry. For proteolysis experiments, materials (0.3 mM with respect to peptide) were pretreated with 0.3 μM thermolysin for approximately 15 h at 37 °C in Tris cleavage buffer, after which the protease was chemically denatured with 10% v/v 0.5 mM EDTA. The reaction was diluted appropriately in competent DMEM to a final concentration of 1 μM polymer and immediately used for cell experiments. Aliquots of the reactions were removed and analyzed on RP-HPLC (25 μL) for the identification of peptide cleavage fragments and TEM for characterization (5 μL). Comparison of the cleavage fragment peak areas to standard curves of the authentic peptide fragment were used to calculate percent cleavage. Controls were prepared by incubating the materials in Tris buffer containing chemically denatured thermolysin (with 10% v/v mM EDTA). Cells were incubated and prepared for flow cytometry analysis as described above. Cell Viability Assay. The CellTiter-Blue fluorescent assay (Promega, cat. #G8081) measures the ability of viable cells to reduce resazurin into a fluorescent product, resarufin. RAW 264.7 cells were plated at a density of 35 000 cells per well of a 96-well plate and allowed to adhere for 24 h. Materials dissolved in DPBS were diluted in DMEM (at a final concentration of 1 μM with respect to polymer) and added to the wells along with a positive control (10% DMSO). Cells were incubated for 24 h at 37 °C. The medium was removed, and cells were washed one time with DPBS (150 μL). Fresh medium (100 μL) without phenol red was added followed by 20 μL of the CellTiter-Blue reagent. Cells were incubated for 3 h prior to measuring fluorescence on a plate reader using 560 nm excitation and 590 nm emission wavelengths. Fluorescence measurements were corrected for background fluorescence of the CellTiter-Blue reagent by subtracting the values of wells containing the reagent in medium in the absence of cells. Percent viability was then calculated by dividing the mean fluorescence of cells treated with material by the mean fluorescence of cells treated with DPBS. At least three technical replicates were performed for each treatment. For all cell experiments, treatment with 10% DMSO displayed approximately 0% cell viability.
Concentrations of solutions were assessed by measuring absorbance of solutions on either a Cary 100 UV−Vis spectrophotometer or nanodrop and comparing to measured extinction coefficients of the materials. DLS and ζ-Potential Characterization. DLS measurements were performed in water using filtered samples (Whatman Puradisk 13 mm 0.2 μm PES membrane) at approximately 0.5 mg/mL for NPs. DLS analysis of kinetic time points from the reaction of thermolysin and P1 was performed on unfiltered samples. For ζ-potential measurements, NPs were dialyzed into 10 mM phosphate buffer at pH 7.5 and measured using a Zetasizer Nano Z (Malvern). It was not practically accessible to measure the ζ-potential of soluble homopolymers because they exhibit even less scattering and require high concentrations to measure their electrophoretic mobilities. All ζpotential values displayed a small range in magnitude, likely due to the low scattering intensities of small, organic-based nanoparticles. Transmission Electron Microscopy. TEM and cryo-EM were performed on a FEI Sphera microscope operating at 200 keV. TEM grids were prepared by depositing small (4 μL) aliquots of sample onto grids (Formvar stabilized with carbon (5−10 nm) on 400 copper mesh, Ted Pella Inc.) that had previously been glow discharged using an Emitech K350 glow discharge unit and plasma-cleaned for 90 s in an E.A. Fischione 1020 unit. The sample grid was rinsed with three drops of water, followed by staining with 1% uranyl acetate solution (again rinsing with three drops). The excess solution was removed by blotting the edge of the grid with filter paper. Micrographs were recorded on a 2k by 2k Gatan CCD camera. Samples for cryo-EM were prepared by depositing 4 μL of sample onto a freshly glow discharged Quantifoil R2/2 TEM grid. The grids were blotted with filter paper under high humidity to create thin films and then rapidly plunged into liquid ethane. The grids were transferred to the microscope under liquid nitrogen and kept at
